[prev in list] [next in list] [prev in thread] [next in thread] 

List:       insight-users
Subject:    Re: [Insight-users] about mean and variance for SimpleFuzzyConnectednessRGBImageFilter
From:       Luis Ibanez <luis.ibanez () kitware ! com>
Date:       2003-12-26 21:23:12
Message-ID: 3FECA6C0.7070708 () kitware ! com
[Download RAW message or body]


Hi Songtao,

You can use the VectorConfidenceConnectedImageFilter
http://www.itk.org/Insight/Doxygen/html/classitk_1_1VectorConfidenceConnectedImageFilter.html


to find an initial segmentation of the object and use
this region for estimating the mean and covariance of
the pixels you want to segment with FuzzyConnectedness.

Since the VectorConfidenceConnected filter will give
you only an initial population for the segmentation,
you can use a small multiplier. Just enough for getting
a reasonable estimation of mean and covariance.

You may want to look at the file:


  Insight/Examples/Segmentation/
               VectorConfidenceConnected.cxx


for an example on how to do this.



Regards,


   Luis


-----------------
宋涛 wrote:

> hi Mr Jin
> 
> Thank you for your last help!
> 
> I have build a segmentation program that uses \
> SimpleFuzzyConnectednessRGBImageFilter. 
> It needs several parameters such as mean and variance.
> 
> But I don't know how to estimate the mean and variance of the object to be used \
> with SetMean, SetVariance. 
> Are there classes that estimate them?? 
> 
> Thanks in advance for your help.
> 
> songtao
> ========================================================
> 网易VIP收费邮箱两周年超值优惠!  �1赠6！！　          http://vip.163.com
> 
> 中国最大的免费邮箱在等你 25兆空间4兆附件！　　　　　 http://mail.163.com
> 
> 点击网易泡泡惊喜无限 全免费手机短信任你发!　　　　　　http://popo.163.com
> "{"�n标鞕ē妜%娝H炄 嗃瑉�"禞+�m��鹈瓛娻?f?X)撸�(!峨rs==


[prev in list] [next in list] [prev in thread] [next in thread]